I’m not sure that I know anyone in the data and analytics field whose platform doesn’t face continuous, or at least annual, budget scrutiny. Analytics is expensive. And when cost-cutting is the order of the day, analytics is a big target. Up shields. Time for another Business Value Inventory.
I’ve spoken with consultants that have completed analytics Business Value Inventories at dozens of different companies, and I’ve done a few myself. Each time we all hope that when we quantify the value we’re generating, the questions will stop. Sometimes they do. At least until next year.
Business Value Inventories typically follow the same pattern: identify a set of active users, send them a survey asking about their use of the analytics environment and the value derived from it, then follow-up to get more details from those that appear to be generating the most value. Aggregate the results and publish.
Business Value Inventories also typically have the same challenge: only a small fraction of interviewees can even hazard a guess at quantifying business value. The most common responses include:
“I don’t know.”
“A lot.”
“I don’t know, but probably a lot.”
“That number should come from my Manager, Director, or Vice President, not me.”
“I couldn’t stand behind any number I gave you.”
Not very useful. It turns out that the numbers they give you aren’t very useful either. At least not on their own. Let’s say that a certain business user claims that they generated $100m in incremental revenue. That’s awesome!! Do you believe it? It seems irresponsible to simply take that number at face value. How exactly was the revenue generated?
We need to approach the business value question a little bit differently, and in a way that makes this exercise much easier for both ourselves and our users.
While very few survey respondents can quantify business value, every single one of them can describe, in detail, what they did with the results of their analyses. The actions taken and / or the decisions made.
This seems to go without saying, and it’s really what you need anyway. The stories. So, ask for them. You can then document assumptions and estimate business value.
Let’s take a very simple example. A staffing analyst is responsible for projecting the number of store check-out clerks to assign each hour. Too many clerks and they stand idle or the customers don’t spend enough time in line looking at the candy displays. Too few and the lines get long and the customers get annoyed (or leave). The staffing analyst optimizes idle time and wait time, while another analyst in another department determines the impact of those metrics across the company. The staffing analyst might not have direct visibility into the impacts, but we can make the association. We can see that idle time decreased by a certain amount with an annual impact of $X, and customer delays decreased by a certain amount with an annual impact of $Y. Benefit is $X + $Y.
These dollar values could not be given a priori, but by focusing on the actions first, the values could be reliably estimated. If someone doesn’t believe the numbers, show them the assumptions and calculations. If they don’t agree with your assumptions, ask them for theirs and recalculate.
—
Unfortunately, Business Value Inventories are only completed occasionally, and with only enough resources to pursue the few stories with the greatest potential benefit. As a result, most overlook a tremendous amount of business value.
Analytics value from individual users is like dark matter in the universe: there’s a whole lot of it out there, it’s largely invisible, and it takes some effort to identify.
But we’ve already maxed out on resources and effort. This means that we need to do something different. Here are three suggestions.
1) Don’t ask for business value. Instead, ask for the business actions driven and the business decisions informed by their queries and analyses.
That would be great to get, but what incentive will the users have to provide any information? In one analytics environment the users were required to briefly describe the purpose for each query they ran. The most common response was just a period. Collecting business actions and decisions should be an ongoing, continuous activity, but maybe not for every query.
The analytics team only has a couple levers. One option is the carrot of granting higher priority to the queries and workloads submitted by those who provide feedback. That works well at first, but since the goal is for everybody to provide feedback, eventually everybody will have high priority which means that nobody will have high priority.
A second approach is less carrot-y but has a greater potential for success: incorporate the collection of business actions and decisions into the standard process for maintaining access to the analytics environment.
2) Require that user access to the analytical environment be renewed every couple of months. Provide the details of their resource consumption during that period and use the collected business actions and decisions as justification.
Quarterly seems to be a good compromise between intruding on every query and waiting so long the users have forgotten what they’ve done.
Even if the answer is “I gave the results to my manager” that’s fine. Follow up by asking, “What did the manager do with the results?” It will take some hand-holding at first, but eventually they’ll get the hang of it and follow the trail themselves.
And putting just a period in the blank doesn’t cut it. If they’re not using the resource, cannot articulate what they did with the results, or didn’t do anything with the results, then their access should be deactivated. A new justification will be required to reestablish it. After all, the company is making an investment in their analytics environment access and it is reasonable to expect a return.
With support and discipline (and some process automation to make it easier for everyone involved), this will become standard operating procedure. Users will become accustomed to thinking about actions and decisions. And telling you about them.
3) Use gamification to drive feedback and compliance.
I’ve often said that “where two or three Vice Presidents are gathered in the presence of a bar chart there will be competition between them.” Keep track of the business benefit and publish scores. Maybe you can’t make the actual numbers public, but perhaps you can abstract the value. Over time you may not even have to request the information. The users may proactively and continually provide it. With their management’s support and encouragement.
Finally, this analysis can be taken further to improve your stewardship of the analytics environment. An individual’s utilization and benefit can be associated with specific resources. Which tools, tables, files, and summaries are being used? Are you expending resources on dormant information assets? A certain table might only be queried once a month, but is the linchpin for an analysis that drives tremendous benefit. Or a bunch of summaries that are never used or have no demonstrable business value might consume a ton of resources.
Raising awareness regarding the true, full benefit of your analytics environment may not totally squelch the noise over its funding and support, but you will always have the information you need for its justification. And data trumps no data.
